Microstructure analysis of martensitic low alloy carbon steel samples subjected to deformation dilatometry

(1)

Contents lists available atScienceDirect

Materials Characterization

journal homepage:www.elsevier.com/locate/matchar

Microstructure analysis of martensitic low alloy carbon steel samples subjected to deformation dilatometry

Jessica Gyhlesten Back

^⁎

, Kumar Babu Surreddi

Materials Technology, Dalarna University, SE-791 88 Falun, Sweden

A R T I C L E I N F O Keywords:

Dilatometry Koistinen-Marburger Martensite Phase transformation EBSD

A B S T R A C T

Low alloy martensitic steels are commonly used in structural and wear resistant applications due to their ex- cellent mechanical properties and abrasion resistance. Martensite phase is generally achieved by rapid cooling, and prior deformation in the austenite region also affects the martensite transformation. It is important to understand the martensite transformation when there is deformation above Ae3. Deformation and quenching simulations have been performed using dilatometry on a low alloy carbon steel. The aim was to determine the influence of deformation above Ae3(prior deformation) on, firstly, the austenite grain size and shape, and secondly, the martensitic microstructure and variant selection. In addition, the hardness of the martensitic structure due to prior deformation has been investigated. The experimental results obtained from electron backscatter diffraction and microhardness tests on the deformation dilatometry test samples were analysed. The orientation relationship Kurdjumov-Sachs has been used to analyse the martensitic variants. The results revealed a deeper understanding of prior austenite grain structure's effect on the martensitic transformation kinetics and its morphology. The martensite laths' misorientation interval 15–48° were used to visualise the prior austenite grain size. The martensitic lath structure is more refined due to increased prior deformation. Shorter martensite formation time promotes a single dominating packet within the prior austenite grain.

1. Introduction

Low alloy martensitic steels are widely used in structural and wear resistant applications due to high hardness and abrasion resistance. The martensitic microstructure is achieved by rapid cooling from the austenitic region to room temperature (RT) or below martensite start temperature (M_s) [1,2]. Uneven deformation at high temperatures (above Ae3) during the manufacturing process and non-homogenous cooling, for example in hot rolled strips, can result in non-uniform residual stresses which can lead to flatness problems. This affects the final mechanical properties, and it is important to control and predict residual stresses in order to ensure reliability. The history of prior deformation and microstructural transformations during the manufacturing process can be helpful to understand the development of residual stresses and to avoid expensive post processing [3]. Most of the hot working processes applies deformation in the austenite region, and the austenite grain size affects the martensitic transformation. It is important to calculate and understand the austenite grain size just before the quenching process (prior austenite). The prior austenite and the resultant martensitic microstructure maintain a crystallographic orientation relationship (OR). Thus, the reconstruction of the prior

austenitic grains from the martensitic structure is possible by data obtained from electron backscatter diffraction analysis (EBSD) [4].

In the present study, low alloy carbon steel samples are subjected to various levels of compressive deformation in the austenitic region by using deformation dilatometry followed by quenching in order to si- mulate the condition of the hot rolled strip. The samples are analysed by EBSD and microhardness. The main aims of the investigation are to determine prior deformation's influence on the martensitic microstructure, the variant selection, and the hardness due to effective plastic straining. This knowledge will improve the understanding of martensitic transformation due to prior deformation in the austenitic region.

2. Theory

Martensite forms by diffusionless transformation when the low alloy carbon steel is cooled rapidly from the austenitic region (athermal transformation) but it can also be generated by plastic deformation [4,5]. The plastic deformation which is an added mechanical energy helps to overcome the energy barrier in order to start the martensitic transformation. Diffusionless transformation means that the carbon content within the crystal is trapped in octahedral sites of a body

https://doi.org/10.1016/j.matchar.2019.109926

Received 30 April 2019; Received in revised form 17 August 2019; Accepted 9 September 2019

⁎Corresponding author.

E-mail addresses:jgb@du.se(J. Gyhlesten Back),kbs@du.se(K.B. Surreddi).

Available online 10 September 2019

T

(2)

centred cubic/tetragonal structure (bcc/bct). The amount of carbon affects the martensitic microstructure which can develop, i.e. lath-, lenticular plate-, thin plate-, and butterfly-martensites [6,7]. They are distinguished by morphology, crystallography or internal structure and their transformation kinetics [8]. The face centred cubic (fcc) structure of austenite can dissolve more carbon than the bcc ferritic structure, however martensite is a ‘stretched’ bcc structure due to the trapped carbon. Lattice parameters of the martensite crystal vary with carbon content and temperature [9]. In low carbon steels (< 0.6 wt% C), the tetragonality of the martensite crystal is almost non-existent [10]. A continuous undercooling drives the martensitic transformation below the Mstemperature. The stresses and the strains which develop during transformation counteract the driving force that the undercooling generates, thus further undercooling is necessary to continue the martensitic transformation. In addition, the undercooling facilitates the growth of the martensitic embryo until it reaches a barrier and continuous undercooling is necessary for it to grow further [2]. However, at the same time further nucleation occurs at other sites where the barrier is lower. This continues until the martensite finish temperature (Mf) is reached and all austenite has transformed to martensite.

Martensite transformation is generally affected by deformation and temperature [11]. The application of external stress at high temperatures in the austenitic region affects the kinetics as it contributes to a mechanical driving force, which either raises or lowers the Ms de- pending on whether the mechanical work aids or opposes the martensitic transformation [12–14]. Prior deformation lowers the austenite grain boundary mobility, thus retarding martensite formation. How- ever, prior deformation also introduces new nucleation sites which advance the formation of martensite. The competing mechanisms, which are decrease in austensite grain boundary mobility and increase in nucleation sites due to prior deformation affect the grain size of prior austenite as well as the martensitic microstructure.

The kinetics of the martensitic transformation are also affected by elastic and plastic strains. The prior austenite and the resultant martensitic microstructure maintain a crystallographic OR due to the ac- commodation effect which handles the volume change as well as shear when fcc transforms to bcc (bct) [8]. The OR is defined as the relation between specific planes and directions of two crystals on either side of the boundary (grain boundary). Thus, the already transformed martensitic crystals strain the remaining adjacent austenite crystals due to the increased volume of the martensite. Therefore, the austenite experiences elastic and plastic strains [4] which in turn affect the martensitic transformation's progress.

During martensite phase transformation, some favoured OR exists between the prior austenite and the product phase which allows the best fit at the interface between the two crystals. In previous studies by Gyhlesten Back et al. [15] of the same low alloy carbon steel the Kurdjumov-Sachs (KeS) OR was found to describe the relation well.

Here, the other possible ORs are mentioned and discussed whether they are suitable or not for the specific steel. However, KeS OR is generally used to explain lath-martensite [16] which develops in low alloy carbon steel, and in the present steel lath-martensite is observed. Therefore, EBSD is used to evaluate the crystallographic orientations of the martensite structure and the KeS OR is used to reconstruct the prior austenite.Table 1presents the common planes and directions of respective phases [17].

There is a strict rule for variant selection where specific combina- tions of two variants (blocks) appear within the prior austenitic grain

[18]. If the laths, in a specific plane are subjected to an external applied stress and thus experiences a positive interaction energy, then the variant is favoured [18]. However, the austenite is almost always constrained during quenching due to internal stresses and temperature gradients in the material, which affect the variant selection in the martensite. Thus, the variant's position and the constraint play a role in variant selection, regardless of the lath's experienced interaction energy due to external stress [19].

A martensite lath can be a single martensitic crystal, called a variant. If the KeS OR is maintained, the symmetry in the cubic system results in 24 possible corresponding crystallographic variants which can be formed within a prior austenitic grain. The austenite fcc crystal (prior austenite grain) have four possible slip planes {111}_γ, which each becomes a martensite packet. The six variants that form on an identical {111}γaustenite plane becomes the possible variants in that packet and are numbered accordingly [18]. Lath boundaries usually have a misorientation of 2–5° to its adjacent laths and the martensite laths (variants) with a similar crystal orientation combine and build blocks.

Packets may consist of several blocks with slightly different orientations in the same habit plane and are misoriented in both low angle and high angle intervals [20,21]. Thus, the lath-martensite has a three-level hierarchy in its morphology, which is laths, blocks, and packets.

The orientations for respective variant denoted by the KeS relationship are plotted on a (001) pole figure inFig. 1and are labelled according to the Morito et al. [22]. In previous work by Gyhlesten Back et al. [15], the KeS OR have been used to reconstruct the prior austenite grains from the martensitic structure in the low alloy martensitic steel samples subjected to heating, austenitisation and quenching. It was determined that the prior austenite grains are best described by the martensite lath boundaries with grain boundary misorientation between 15 and 48°. However, the present study is on samples deformed at high temperature above A_e3followed by quenching to RT and the resultant martensite will be affected by the prior deformation. The prior deformation will affect (i) the size of laths, blocks and packets, (ii) the final strength of the sample [23] and (iii) perhaps also the lath boundaries misorientation to adjacent laths. The experimental setup using deformation dilatometry combined with microstructure analysis will increase the knowledge about the correlation between austenite

Table 1

Kurdjumov-Sachs orientation relationship between the austenitic and martensitic crystals.

OR Plane Direction

K-S {111}γ//{110}α 〈110〉γ//〈111〉α

(001)

100

_γ

010

_γ

γ

11

_γ

γ

00

_γ

01

_γ

19 22

23 20

21 24

9 12 11

8 7 10

14 17 13

16 15

18 5 2

6 4 3

1

Fig. 1. (001) pole figure of the 24 martensitic variants (indicated by numbers and colouring with respect to habit plane) generated from a single prior austenite crystal that maintain the Kurdjumov-Sachs orientation relationship.

(3)

and martensite.

3. Experimental methods

The material used in the present study is a low alloy carbon steel and the composition can be found inTable 2. The samples were cy- lindrical with the flat edge towards the rolling direction and they had a size of 5 mm in diameter and 10 mm in length. The samples have been taken from a strip which has been hot rolled to a thickness of 6 mm and has thereafter been quenched to a fully martensitic structure. The strip was coiled at a temperature below 100 °C. Deformation dilatometry experiments were performed using a differential deformation dilatometry using Bähr DIL 805 A/D at Swerea KIMAB in Sweden. The sample was held between two push rods made of Al2O3and the deformation was controlled by hydraulic pressure. The dilatation was measured in relation to the original length of the sample with an accuracy of 50 nm.

Before inserting the sample, thermocouples were spot-welded onto the sample's surface and its initial length was measured. The heating of the sample was done by induction heating at 10⁻⁴bar vacuum, and the cooling was done by high volume flow of helium gas.

The heating-deformation scheme were, heating to 890 °C with 10 °C/s and held for 300 s for homogenous austenitisation. Thereafter the samples were subjected to 0, 5, 10, 20, 30 and 40% compression ratio using compression rate 10 mm/s. They were then cooled down to room temperature (RT) by 100 °C/s which previously in Gyhlesten Back et al. [15] has been determined to yield a similar martensitic structure as the conventional hot rolling and quenching process.

The cross sections of the quenched samples were prepared by fol- lowing conventional metallographical sample preparation and then etched with 3% Nital. The samples were investigated by scanning electron microscopy (SEM) using electron backscatter diffraction (EBSD) and also Vickers hardness measurements. The analysed surface was the specimen's center (half the length) and 1/4 from the center in radial direction. The EBSD analysis was performed using Zeiss Ultra 55, field emission gun scanning electron microscope (Carl Zeiss Microscopy, Jena, Germany). Data acquisition and post-processing of crystalline data obtained from EBSD were performed using HKL Channel 5 software (version 5.12.61.0, Oxford Instrument NanoAnalysis, Oxford, UK). The samples were prepared according to EBSD analysis requirements. See the EBSD acquisition parameters in Table 3. A data cleaning was performed prior to the EBSD analysis in Channel 5. The cleaning included replacement of incorrectly indexed isolated points, and filling of unindexed point which had at least 4 indexed neighbors, with copies of neighboring points. The band contrast

and IPF colouring was applied to the EBSD micrographs in addition to the grain boundary markings to enhance the interpretation of the martensitic structure of the samples. The crystallographic data was further processed using KeS OR (in-house Matlab model) to calculate the prior austenite from the martensitic structure of the samples [15]

and the results were further used to create band contrast and frequency plots. The Vickers hardness measurements were performed in a Micro- Combi tester from CSM Instruments with a load of 500 gf with 15 s dwell time.

4. Results and discussion

The prior austenite grain structure has an impact on the resultant martensite transformation and microstructure, such as the necessary undercooling for the martensite transformation to start, martensite lath size and the variants selection [15]. The misorientation profiles of the martensitic microstructure of each deformation dilatometry samples was analysed from the EBSD results. The misorientation interval is a proper tool to use when visualising the prior austenitic grain boundaries, and the grains' sizes and shapes in deformed austenite. Fig. 2 show the EBSD micrographs of the samples after (a) 0%, (b) 20%, (c) 30% and (d) 40% compression. The band contrast, inverse polar figure (IPF) colouring, and grain boundary misorientation from 15 to 48°

(black lines) are applied to the EBSD images in order to comprehend the prior austenite grain structure. The IPF colour components paint the pixels based on the orientation of the lattice. Reference direction for the present results was Z axis (IPF Z). The IPF colouring enhances the interpretation of the angle differences between the martensitic laths and also between the prior austenite grains.Fig. 2(a)–(d) illustrate the decrease in grain size by increasing compression ratio, and the change in morphology of the martensitic structure. Larger grains facilitate the martensite transformation and smaller grains reduce the growth of martensite, which leads to finer laths [24] as observed inFig. 2.

The average prior austenite grain size is measured from the EBSD micrographs for all the samples using the linear intercept method and the results are plotted inFig. 3(a). The microhardness results of all samples are plotted and shown inFig. 3(b). The average grain size decreases with increasing compression ratio. During plastic deformation, structural heterogeneities such as deformation bands, twins and high angle grain boundaries, serve as nucleation sites. In order to fulfill the forming compatibility, when grain size decreases (above Ae3) the misorientation increases. Also, a higher density of geometrically necessary dislocations is stored per unit volume to achieve strain gradients at the grain boundaries, which is a result of the higher amount of grains. Nucleation of new grains is strongly dependent on grain size.

Smaller grains increase the stored energy available for nucleation, which increases the driving force for dynamic recrystallization. The increase of available nucleation sites by increasing the grain boundary area per unit volume cause a change in the dynamic recrystallization behavior [25].

The difference in prior austenite grain size between 0% and 5%

compression is very small and this corroborates with the small decrease in microhardness. The error bar for 20% compression ratio show large variations in grain size, which also indicates large variations in the martensite lath structure. In both these cases variants selection play a role and affect the local hardness, seeFig. 3(b). From 10% compression level and above, the grain size decrease and the hardness increase with increasing prior deformation.

There is a correlation between the martensitic variants formed in the samples and the resultant microhardness. The local differences and distributions of the martensite variants affect the local hardness. The frequency of the martensite variants is generally high around misorientation 60°, thus these variants are favoured. Martensitic variants with misorientation angle 60° to the first variant (V1) are V2, V3 and V5, and Table 4 shows their crystal orientations. The frequency of martensite boundaries with the misorientation angles for all samples is Table 2

The composition of the low alloy carbon steel.

Element C Si Mn P S Cr Ni

[Wt%] 0.176 0.244 1.27 0.009 0.001 0.25 0.04

Mo Al Nb V Ti N B

[Wt%] 0.017 0.038 0.001 0.008 0.01 0.002 0.004

Table 3

Settings for the acquisition of data during the EBSD analysis.

Accelerating voltage 15 kV

Aperture size 120 μm

Working distance ~15 mm

Step size 0.25 μm

EBSD camera binning mode 4 × 4

Hough resolution 60

Phases for acquisition iron fcc, iron bcc

Band detection mode edges

Number of bands detected 8

Number of reflectors 32 fcc, 43 bcc

(4)

plotted inFig. 4. A clear variation in frequency can be observed at low misorientation angles (0–20°) and high misorientation angles (40–62°) as shown inFig. 4(a) and b) respectively. The frequency of low angle martensitic boundaries generally increases and high angle martensitic boundaries decreases with increasing prior deformation. The peaks around 60° inFig. 4(b) clearly show a decreasing trend with increasing prior deformation, however, the samples with 5 and 20% compression do not follow the general pattern. The lowest measured hardness for prior deformation by 5% is probably due to the martensitic variant

selection and the balance between developed low and high angle martensitic boundaries. The experienced interaction energy results in softer martensite lath structure in 5% prior deformation as compared to the undeformed sample. Low angle martensitic boundaries are generally increasing with increasing prior deformation as shown in Fig. 4(a), which results in a finer martensitic structure and leads to increased microhardness. According to the Hall-Petch relation, the yield stress and hardness increases with decreasing grain size. In martensitic steels as shown by Morito et al., the yield stress and hardness also Fig. 2. EBSD micrographs of the martensitic microstructure after prior deformation by (a) 0%, (b) 20%, (c) 30% and (d) 40%. Band contrast, IPF colouring (IPF-Z0) and marked prior austenitic boundaries (black) are applied to the EBSD images.

0 5 10 15 20 25

-5 0 5 10 15 20 25 30 35 40 45

Prior austenite grain size [μm]

Compression ratio [%]

a

420 430 440 450 460 470 480 490

-5 0 5 10 15 20 25 30 35 40 45

HV0.5

Compression ratio [%]

b

Fig. 3. (a) Prior austenitic grain size and (b) microhardness variation as a function of compression ratio.

(5)

depends on packet and block size [26]. Hidalgo et al. [27] found that the dislocation density affects the strength of martensitic steels. The dislocation density increases by increasing amount of prior deformation and also the number of lath, block and packet boundaries where the dislocations need to pass or break through increases [27]. Therefore all martensitic boundaries contribute to yield stress and hardness in martensitic steels.

Further microstructure analysis was carried out using the results from the EBSD analysis. In order to confirm the misorientation interval 15–48° as a prior austenite grain boundary, a comparative analysis was made using EBSD band contrast images applying KeS OR.

The prior austenite grain boundaries are marked according to the KeS OR relationship as shown inFig. 5(a), (c), (e), and (g), however some of the martensite lath boundaries are included as austenite boundaries. The misorientation interval 15–48° works very well for the deformed samples and is more accurate than the KeS OR, seeFig. 5(b), (d), (f) and (h). Increasing the misorientation interval to 49° and 50°, revealed more lath boundaries and the visibility of the prior austenitic boundaries became poor.

By viewingFigs. 2 and 5, it is clear that the prior austenite grain size decreased with an increasing amount of prior deformation. There was a pronounced refinement of the grain size above compression ratio 30%.

Smaller prior austenite grains experience shorter time for martensite formation, which means that the packet that coincide with the strain dominate the grain [28]. Similar findings of grain refinement have been found by Gomez et al. [29] during hot rolling. Their focus was recrystallization between the interpasses and its effect on grain size and shape. At lower temperatures, the recrystallization was delayed resulting in an uneven distribution of fully recrystallized grains and cer- tain areas showed elongated grains due to the absence of recrystallization. The stop recrystallization temperature for the investigated steel was 897 °C. For the low alloy carbon steel used in the present deformation dilatometry experiments, there are no elements which can retard the recrystallization, only time is of the essence. In the work by Engberg et al. [30], the time from deformation to initiation of

recrystallization was measured as < 1 s for a steel deformed by 40% at 900 °C. Since the quenching in the present experiments took place di- rectly after the deformation, the time for static recrystallization was short. However, dynamic recrystallization seemed to take place in the specimens deformed by higher compression ratios, since there were areas with very small grains which were not randomly distributed already by 30% compression ratio. Increasing amount of prior deformation increases the stored energy which can be used for dynamic recrystallization.

Further, the shape of the prior austenite grains changed between deformation 30 and 40%. The prior austenitic boundaries moved from clear and straight to irregular curved boundaries. These grains appear in periodic bands which reflect the forging cross seen in the specimens that were heavily bulged due to the high compression ratio, see Fig. 6(a). In the samples deformed by 20%, the prior austenite grains' shapes are the same as for 0, 5 and 10% deformation, but the lath structure is changed. The laths became more elongated and more aligned towards the same direction. There is a clear trend that the low angle martensite boundaries are increasing by increasing amount of strain. Similar results were gained in the work of Zolotarevskii et al.

[31].

In the specimen deformed by 40%, the band contrast and the marked prior austenite boundaries show variations between the grains as shown inFig. 6(b) and (c). Some of the prior austenite grains with curved or irregular boundaries generally show none or very little misorientation within the grains. They appear to have no martensite lath structure. These grains generally occur in periodic bands as shown in Fig. 6(a). From the EBSD results, they were identified as bcc structure only and can therefore be interpreted to be ferritic grains. It has previously been found that ferrite formation during rapid cooling is pro- moted by (i) low carbon content, (ii) heavy deformation of the austenite and (iii) by refinement of austenite close to the Ae3temperature. All these factors reduce the hardenability of austenite which in turn facilitate the nucleation and growth of ferrite. Austenite deformed by 40%

at temperatures close to Ae3 increased the rate of the γ to α Table 4

Crystal orientations of variants V2, V3 and V5 from the Kurdjumov-Sachs orientation relationship and crystallographic angle/axis misorientation from V1 [15].

Variant Austenite OR K-S Martensite OR K-S Misorientation angle from V1 [°] Misorientation axis from V1

V2 (111)[−101] (011)[−111] 60.0 0.577 −0.577 0.577

V3 (111)[01−1] (011)[−1−11] 60.0 0 −0.707 −0.707

V5 (111) [1−10] (011)[−1−11] 60.0 0 0.707 0.707

0 5 10 15 20 25

0 5 10 15 20

Frequency of found martensite boundaries [%]

Misorientation [°]

0%

5%

10%

20%

30%

40%

a

0 5 10 15 20 25

40 45 50 55 60 65

Misorientation [°]

0%

5%

10%

20%

30%

40%

b

0 5 10 15 20 25

0 5 10 15 20

Misorientation [°]

0%

5%

10%

20%

30%

40%

a

0 5 10 15 20 25

40 45 50 55 60 65

Misorientation [°]

0%

5%

10%

20%

30%

40%

b

Fig. 4. The distribution of martensite boundaries with the misorientation 0–20° (a) and 40–62° (b) for undeformed and deformed samples. The frequency plots are not normalised.

(6)

(caption on next page)

(7)

transformation during continuous cooling [32]. Further, the alloying element niobium assists the retardation of austenite recrystallization which refined the grain size but also increased nuclei for the austenite to ferrite transformation (γ → α) [33]. The ferrite nose in the continuous cooling temperature diagram advanced to shorter holding time when the grain size decreased. Thus, it is probable that the activation energy for ferrite formation was lowered in these specific bands due to the strained austenite.

The morphology of the grains changes by increasing the amount of prior deformation. The grains change already at 10% but there is a clear change by 20% prior deformation. The laths became flattened and more elongated in a rather ordered system in the 20% deformed samples. The change in morphology clearly shows that refinement of the austenitic grain size by increasing the amount of prior deformation affects the characteristics of the martensitic transformation [27]. This is due to the altered stress state which affects the martensite transformation kinetics [13]. From this perspective it is interesting to know the differences, in the variants selection and the resulting martensite boundaries, in undeformed and severe priory deformed samples.

Fig. 7 presents a comparison of the variants and martensite boundaries for no deformation and 40% deformation. The martensitic boundaries which have high angle misorientation are decreasing in number by increasing the amount of deformation while the low angle misorientations are increasing as observed inFig. 7(a). The same trend is found for the frequency of martensitic variants, seeFig. 7(b). In the sample which is not subjected to deformation, the variants with misorientation 60° are most frequent while the variants with lower misorientation angles increase in the 40% deformed sample. The variant selection due to prior deformation also affects the block and packet boundaries, which is shown inFig. 8.

The EBSD images in Fig. 2 show large variations of IPF colours within the grains in the undeformed sample as compared to 40% deformed sample. This can be correlated to the specific variants selection i.e., laths, blocks and packets.Fig. 8shows magnified IPF images of the prior austenitic grains subjected to compression ratio (a) 0% and (b) 40%. Fig. 8(a) shows that undeformed austenite gives a martensitic structure with several packets containing almost all kinds of variants.

However, deformed austenite reveals one dominant packet with variants V19 toV24 (denoted by I inFig. 8(b)) and another smaller packet with variants V13-V18 (denoted by II in Fig. 8(b)), in which several smaller blocks can be observed.

Previous studies have shown that block boundaries restrict dislocation motion and have significant impact on the steel's strength.

Some grains also contain sub-blocks which are formed in highly deformed samples, which do not contribute to the strengthening in the same amount as the conventional block boundaries [34]. Morito et al.

[28] also found that the number of the various variants which can appear in lath martensite decrease when the grain size is decreased.

Further, the packet size, which is considered as the effective strengthening grain size in martensitic structures, decreases. However in Ta- mura et al. [35], they observed the opposite i.e. the packet size increases and block size decreases with increasing prior deformation, which confirms the theory from Mine et al. [34] that the blocks affect dislocation movement. The present experiments and analysis show that low angle martensite boundaries increase due to prior deformation and high angle martensite boundaries decrease due to packet growth, see Figs. 7 and 8.

InFig. 9 {100} pole figures of two martensitic samples, (a) 0%

deformed sample and (b) 40% deformed sample, are shown in order to compare orientations of the lath structure. The pole figures were plotted using two different contouring, where the left ones show the intensity of the lath's misorientation and the right ones show the lath's direction, see the IPF Z legend.

In general, the texture of deformed austenite follows the compression's direction (slip plane and direction). However, due to nucleation and formation of new equiaxed grains, the texture is weak in both the undeformed and the heavy deformed samples [36]. InFig. 9, it is observed that the samples' textures goes from (101) towards (001) after 40% prior deformation before quenching to martensite. The martensitic structure is mainly affected by the grain size and residual strains and stresses, which affect the variant selection [36].

5. Conclusion

The conclusions of the present work are summarised as follows:

•

The prior austenite grain boundaries were identified from the martensitic microstructure using misorientation interval 15–48° and reconstruction by KeS OR. It clearly shows that the austenite grain size decreases by increasing prior deformation.

•

In general, the microhardness in the samples increased by increasing the amount of prior deformation and the local differences and Fig. 5. [1–36]Comparison of the prior austenitic grain boundaries for the samples subjected to various compression ratio using the Kurdjumov-Sachs orientation relationship (a), (c), (e), and (g) and pre-set misorientation interval 15–48° (b), (d), (f) and (h).

Fig. 6. Band contrast of a 40% deformed sample (a), band contrast image from the selected area (b), and IPF image with grain boundaries (red) of the selected area (c). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

(8)

distributions of the martensite variants formed in the samples affect the local hardness and also the microhardness results.

•

The martensitic formation characteristics change in deformed austenite compared to undeformed austenite. This is due to the altered stress state which affects the martensite transformation kinetics.

•

The frequency of low angle martensitic boundaries generally increases and high angle martensitic boundaries decreases with increasing prior deformation.

•

The smaller prior austenite grains affect the martensite laths' growth. The laths collide and stop growing sooner and therefor the laths become thinner and shorter.

•

Heavy deformation causes bulging of the sample which results in a banded structure in the sample with 40% prior deformation. In the shear bands the austenite is very strained, thus ferrite growth profits

and develops in advance of martensite growth in these areas.

Data availability

The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study.

Acknowledgements

This work has been conducted within the framework of a PhD project in collaboration with SSAB EMEA and Dalarna University in Sweden. The deformation dilatometry experiments were performed at Swerea KIMAB in Stockholm and the remainder of the tests and analysis were conducted at Dalarna University and SSAB in Borlänge. The 0

5 10 15 20 25

10.5 12.5 14.5 16.5 18.5 20.5 22.5 24.5 26.5 28.5 30.5 32.5 34.5 36.5 38.5 40.5 42.5 44.5 46.5 48.5 50.5 52.5 54.5 56.5 58.5 60.5 62.5

Frequency of martensite boundaries [%]

Misorientation [°]

0%

40%

a

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

10.5 11.5 12.5 13.5 14.5 15.5 16.5 17.5 18.5 19.5 20.5

Frequency of martensite boundaries [%]

Misorientation [°]

0%

40%

0 10 20 30 40 50 60

10.53 14.88 20.61 21.06 47.11 49.47 50.51 51.73 57.21 60

Frequency of variants [%]

Misorientation [°]

0%

40%

b

Fig. 7. The frequency of martensite boundaries and their misorientation distribution for compression ratio 0 and 40% (a) and the variant frequency with mis- orientation (b). The lower misorientation angle in (a) marked with red dashed box is magnified and presented in the inset in (a). The both frequency plots are not normalised. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 8. Magnified IPF images with grain boundaries (red) of 0% deformation sample (a) and 40% deformation sample (b). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

(9)

authors acknowledge SSAB EMEA for their contribution with materials and financial support.

Special thanks to sponsors of the Steel Industries Graduate School which are: Regional Development Council of Dalarna, Regional Development Council of Gävleborg, County Administrative Board of Gävleborg, The Swedish Steel Producers' Association, Dalarna University, Sandviken Municipality and SSAB EMEA.

References

[1] J.R. Davis, Metals Handbook Desk Edition, 2nd edition, ASM International, 1998.

[2] L. Kaufman, M. Cohen, Thermodynamics and kinetics of martensitic transformations, Prog. Met. Phys. 7 (1958) 165–246.

[3] S. Hashmi, G.F. Batalha, C.J. Van Tyne, B.S. Yilbas, Comprehensive materials processing - 13 volume set, Advanced Forming Technologies, 3 Elsevier, 2014.

[4] H.K.D.H. Bhadeshia, Worked Examples in the Geometry of Crystals, Institute of metals, 2006.

[5] M. Naghizadeh, H. Mirzadeh, Microstructural evolutions during annealing of plas- tically deformed AISI 304 austenitic stainless steel: Martensite reversion, grain refinement, recrystallization, and grain growth, Metall. Mater. Trans. A 47 (8) (2016) 4210–4216.

[6] T. Maki, Microstructure and mechanical behaviour of ferrous martensite, Mater. Sci.

Forum 56–58 (1990) 157–168.

[7] A. Shibata, S. Morito, T. Furuhara, T. Maki, Substructures of lenticular martensites with different start temperatures in ferrous alloys, Acta Materiala 57 (2009) 483–492.

[8] J.W. Christian, Physical Properties of Martensite and Bainite: Military Transformations—An Introductory Survey, Humphries & Co Ltd, London and Bradford, 1965.

[9] R. Abbaschian and R.E. Abbaschian L. and. Reed-Hill. Physical Metallurgy Principles. Cengage Learning, Stamford, USA, 4 edition, 2009.

[10] B. Hutchinson, J. Hagström, O. Karlsson, D. Lindell, M. Tornberg, F. Lindberg, M. Thuvander, Microstructures and hardness of as-quenched martensites (0.1–0.5%

c), Acta Mater. 59 (2011) 5845–5858.

[11] E. Pereloma, D.V. Edmonds, Phase transformations in steels, Diffusionless Transformations, High Strength Steels, Modelling and Advanced Analytical Techniques, volume 2, Woodhead Publishing, 2012.

[12] J.R. Patel, M. Cohen, Criterion for the action of applied stress in the martensitic

transformation, Acta Metall. 1 (5) (1953) 531–538.

[13] S. Denis, E. Gautier, A. Simon, G. Beck, Stress-phase-transformation interactions - basic principles, modelling, and calculation of internal stresses, Mater. Sci. Technol.

1 (10) (1985) 805–814.

[14] D.A. Porter, K.E. Easterling, Phase Transformations in Metals and Alloys, Chapman

& Hall, London, 1992.

[15] J. Gyhlesten Back, G. Engberg, Investigation of parent austenite grains from martensite structure using EBSD in a wear resistant steel, Materials 10 (5) (2017) 453.

[16] M. Hillert, J. Ågren, A. Borgenstam, Mikro och nanostrukturer i materialdesign.

Kungliga Tekniska Högskolan, (2005).

[17] P.P. Suikkanen, C. Cayron, A.J. DeArdo, L.P. Karjalainen, Crystallographic analysis of martensite in 0.2C-2.0Mn-1.5Si-0.6Cr steel using EBSD, Journal of Materials Science & Technology 27 (10) (2011) 920–930.

[18] H. Kitahara, R. Ueji, N. Tsuji, Y. Minamino, Crystallographic features of lath martensite in low-carbon steel, Acta Mater. 54 (5) (2006) 1279–1288.

[19] S. Kundu, A.K. Verma, V. Sharma, Quantitative analysis of variant selection for displacive transformations under stress, Metall. Mater. Trans. A 43 (7) (2012) 2552–2565.

[20] S. Morito, Y. Adachi, T. Ohba, Morphology and crystallography of sub-blocks in ultra-low carbon lath martensite steel, Mater. Trans. 50 (8) (2009) 1919–1923.

[21] S. Morito, H. Tanaka, R. Konishi, T. Furuhara, T. Maki, The morphology and crystallography of lath martensite in Fe-C alloys, Acta Mater. 51 (6) (2003) 1789–1799.

[22] G. Miyamoto, N. Iwata, N. Takayama, T. Furuhara, Quantitative analysis of variant selection in ausformed lath martensite, Acta Mater. 60 (3) (2012) 1139–1148.

[23] D.V. Doane, J.S. Kirkaldy (Eds.), Hardenability Concepts with Applications to Steel, Metallurgical Society of AIME, 1978.

[24] H. Gungunes, E. Yasar, T.N. Durlu, The effect of austenitizing time on martensite morphologies and magnetic properties of martensite in Fe-24.5%Ni-4.5%Si alloy, J.

Mater. Sci. 42 (15) (2007) 6102–6107.

[25] C. Rehrla, S. Kleberb, O. Renka, R. Pippan, Effect of grain size in compression deformation on the microstructural evolution of an austenitic stainless steel, Mater.

Sci. Eng. A 540 (2012) 55–62.

[26] S. Morito, H. Yoshida, T. Maki, X. Huang, Effect of block size on the strength of lath martensite in low carbon steels, Mater. Sci. Eng. A 438–440 (2006) 237–240 (Proceedings of the International Conference on Martensitic Transformations).

[27] J. Hidalgo, M.J. Santofimia, Effect of prior austenite grain size refinement by thermal cycling on the microstructural features of as-quenched lath martensite, Metall. Mater. Trans. A 47 (11) (2016) 5288–5301 November.

[28] S. Morito, H. Saito, T. Ogawa, T. Furuhara, T. Maki, Effect of austenite grain size on the morphology and crystallography of lath martensite in low carbon steels, ISIJ Int.

Fig. 9. The pole figures, {100}, of two martensitic samples, with 0% deformation a) and with 40% deformation b).

(10)

45 (1) (2005) 91–94.

[29] M. Gomez, L. Rancel, B.J. Fernandez, S.F. Medina, Evolution of austenite static recrystallization and grain size during hot rolling of a v-microalloyed steel, Mater.

Sci. Eng. A 501 (1–2) (2009) 188–196.

[30] G. Engberg, L. Lissel, A physically based microstructure model for predicting the microstructural evolution of a C-Mn steel during and after hot deformation, Steel Research International 79 (1) (2008) 47–58.

[31] N.Y. Zolotarevskii, A.A. Zisman, S.N. Panpurin, Y.F. Titovets, S.A. Golosienko, E.I. Khlusova, Effect of the grain size and deformation substructure of austenite on the crystal geometry of bainite and martensite in low-carbon steels, Metal Science and Heat Treatment 55 (9–10) (2014) 550–558 January.

[32] N. Fonstein, Advanced High Strength Sheet Steels: Physical Metallurgy, Design, Processing, and Properties, Springer International Publishing, Switzerland, 2015.

[33] Y. Granbom, Structure and Mechanical Properties of Dual Phase Steels - An Experimental and Theoretical Analysis, Phd thesis Royal Institute of Technology School of Industrial Engineering and Management Materials Science and Engineering Division of Mechanical Metallurgy, 2010.

[34] Y. Mine, K. Hirashita, H. Takashima, H. Matsuda, K. Takashima, Micro-tension behaviour of lath martensite structures of carbon steel, Mater. Sci. Eng. A 560 (2013) 535–544.

[35] I. Tamura, K. Tsuzaki, T. Maki, Morphology of lath martensite formed from deformed austenite in 18%Ni maraging steel, J. Phys. 43 (12) (1982) 551–556.

[36] N.F. Viana, C. dos Santos Nunes, H.F.G. de Abreu, The variant selection in the transformation from austenite to martensite in samples of maraging-350 steel, Journal of Materials Research and Technology 2 (4) (2013) 298–302.